Re: A bad behavior under autocommit off mode

Hiroshi Inoue wrote:
> Bruce Momjian wrote:
> >
> > OK, I have a patch to fix this bug. The basic problem is that when a
> > multi-query string is passed to the backend, it is treated as a single
> > transaction _unless_ a transaction or GUC command appears in the string.
> > When they appear, a transaction is forced, but the normal transaction
> > state machine has been bypassed, meaning in:
> >
> > SET autocommit TO off; SELECT 1; COMMIT;
> >
> > when the COMMIT arrives, the transaction state machines hasn't seen the
> > SELECT because the mechanism is bypassing the state machine to try and
> > get everything into the same transaction.
> >
> > This patch removes that "stuff all queries into a single transaction"
> > behavior and makes them function just like queries arriving separately.
> > This does BREAK BACKWARD COMPATIBILITY. However, if they want the old
> > behavior, they just need to wrap BEGIN/COMMIT around the query string.
>
> Does the change worth the trouble ?
> Please don't break BACKWARD COMPATIBILITY easily.

It clearly fixes an existing bug, and I asked on general to see if
anyone has any problem with the change. My guess is that more people
are surprised by the group-string-as-a-single-transaction as people who
use the feature, so I see it as the removal of surprising functionality.

We will mention it in the release notes, and I can even supply a patch
for those who want it kept. In fact, I can easily make it a compile
option --- the change is only a single conditional test.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073